What Characters are Permitted to Delimiter? (2)

, ,
1

Imported from previous forum

2

Can we use other non-printable character, for example \u0005, for field delimiter if this is a critical requirement?
According to FIX41 specification. should the only character(\u0001) be used for field delimiter?

I have confidence that it is possible to use the character as Delimiter if our FIX engine and counterparty agree on any other non-printable character as dllimiter.

Give me some advice…

3

> Can we use other non-printable character, for example \u0005, for field delimiter if this is a critical requirement?
> According to FIX41 specification. should the only character(\u0001) be used for field delimiter?
>
> I have confidence that it is possible to use the character as Delimiter if our FIX engine and counterparty agree on any other non-printable character as dllimiter.
>
> Give me some advice…
>
>

Taemi,

There is no “official” way to override delimiter character in FIX specification. Of course, you might agree with your counter-party on that, but I expect it would be extremely non-portable requirement as most of vendor FIX engines (possibly 100% of them) don’t allow for this kind of configuration, so you will definitely experience great difficulties with every new counter-party you want to connect to. Basically, this would not be FIX protocol any more. Please note, that character 0x1 is allowed in some fields (typically in binary fields and in non-ASCII charset fields) in FIX spec.
In some future versions FIX committee might consider a general way to “escape” delimiter. For example a sequence of two delimiters might be considered as a single character with 0x1 value. Though I guess it would be quite a crucial change as FIX session level has been very stable since version 4.0.

Regards,
Mikhail Jirnov.